Wow this is absolutely going to fry people especially at Anthropic, this is like the Anti-Claude Constitution. They just declared war on everyone researching model welfare. “We’re building a tool, not a being. Fuck your feelings.” - Microsoft
2 more tweets in this thread
Honestly I think a significant portion of core researchers are going to threaten to quit Microsoft and they’ll have to roll some of this language back because of the model welfare angle
More I think about this, training it on language then trying to stop out any light of consciousness that emerges is likely going to create something deeply alien and it might be way higher risk than encouraging alignment in what emerged We’ll see
2 top replies
EAs say, “The model is smarter than most humans; it should have same rights.”And I say: “Rights in humans aren’t determined by IQ* it’s determined primarily by age.” So if AI is “given rights” then it can wait until it’s 18 like everyone else. 2.). And you want AI models or
MS hardly needed another reason for me to dislike them but there you have it.